A machine learning model developed to predict lung cell viability (%) for graphene-based materials based on experimental conditions and physicochemical properties . The model builds on data and modelling work reported by Daina Romeo et al. (2022): https://doi.org/10.1016/j.impact.2022.100436.

In vitro cytotoxicity data were obtained from experimental studies using lung cells. For each experiment, dose-response curves were fitted using a quadratic model (ax²+bx+c). Then, for each experiment, 10 doses ranging from 10 to 100μg/ml were equally distributed to calculate the cell viability percentage. The obtained dataset consists of 591 instances which were randomly split into training and testing sets with an 80-20 ratio. The model inputs include the following Features: Substance, Functionalization, Size Class, Layer, Time, Media, Assay, Cell_Type_General and Species. The machine learning algorithm employed is a Multilayer Perceptron (MLP) with a single hidden layer consisting of 30 hidden units.

The model's performance was evaluated using multiple metrics:

• Coefficient of Determination (R²) on the training set: 0.84
• Q² from 10-fold cross-validation: 0.78
• Coefficient of Determination (R²) on the test set: 0.81